Tek-Tips is the largest IT community on the Internet today!
Members share and learn making Tek-Tips Forums the best source of peer-reviewed technical information on the Internet!
-
Congratulations TugboatEng on being selected by the Eng-Tips community for having the most helpful posts in the forums last week. Way to Go!
RPM sensor motor connected to VFD 2
- Thread starter ePOWEReng
- Start date
Skogsgurra
Electrical
I find that question a bit "backwards". Are you thinking of an application and wondering if you can use a VFD without a speed sensor?
Anyhow, a paper machine drive almost certainly cannou work without a speed transducer. Today's specifications say at least 0.05 and usually 0.01 percent speed accuracy. That is something you cannot have without a speed transducer, often an incremental encoder with around 1024 PPR.
A test bench for motors used to run without speed transducers and you could have something like +/- 0.5 percent accuracy from the speed estimator in a VFD. But that is not enough any more. Efficiency needs to measured with at least 0.1 percent accuracy and since the torqu transducer is difficult to get any better, the speed measurement needs to be as good as it can be.
Generally, if you need better than around 0.5 percent accuracy over torque, speed and temperature range, you need a speed transducer.
Some built-in estimators/observers can do better than that if calibrated continuoiusly. But that is usually only in special situations.
The SKF Sensor Bearing is a good alternative to a "real" encoder. It is built into the bearing and is thus very well protected. It is also low cost and can achieve high accuracy if reciprocal speed measurement is used.
Gunnar Englund
--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...
Anyhow, a paper machine drive almost certainly cannou work without a speed transducer. Today's specifications say at least 0.05 and usually 0.01 percent speed accuracy. That is something you cannot have without a speed transducer, often an incremental encoder with around 1024 PPR.
A test bench for motors used to run without speed transducers and you could have something like +/- 0.5 percent accuracy from the speed estimator in a VFD. But that is not enough any more. Efficiency needs to measured with at least 0.1 percent accuracy and since the torqu transducer is difficult to get any better, the speed measurement needs to be as good as it can be.
Generally, if you need better than around 0.5 percent accuracy over torque, speed and temperature range, you need a speed transducer.
Some built-in estimators/observers can do better than that if calibrated continuoiusly. But that is usually only in special situations.
The SKF Sensor Bearing is a good alternative to a "real" encoder. It is built into the bearing and is thus very well protected. It is also low cost and can achieve high accuracy if reciprocal speed measurement is used.
Gunnar Englund
--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...
-
1
- #3
-
1
- #4
When talking about speed control and error in a drive/motor system, it is important to understand that there are several different aspects to the issue.
First, speed error is generally due to changes in torque demand. In an induction motor, this error is mostly slip. So the question becomes, how well does the drive compensate for torque induced slip speed changes. With a good vector drive, this can get down in the range of one-tenth of motor slip without an encoder. If you need better than that, an encoder is required. Note here that the error is a result of torque changes. If your torque doesn't change, you won't have much speed error to start with.
Second, in some applications, especially those involving web products and tension control, cumulative error is just as important as actual error. For example, even if you are very accurate with actual error, if it is all negative or all positive, eventually you are going to have too much or too little tension. No encoderless system will assure non-cumulative error. For that you need an encoder.
Third, speed reference error is often overlooked. That is error either in the speed signal going into the drive or error in the drive translating the input command into an actual output speed. Usually, the majority of this error is due to the analog input terminal analog-to-digital conversion. A 10 bit resolution A/D input will not be nearly as accurate as a 14 bit resolution input. This is a matter of purchasing a drive with the input resolution adequate for the intended purpose.
First, speed error is generally due to changes in torque demand. In an induction motor, this error is mostly slip. So the question becomes, how well does the drive compensate for torque induced slip speed changes. With a good vector drive, this can get down in the range of one-tenth of motor slip without an encoder. If you need better than that, an encoder is required. Note here that the error is a result of torque changes. If your torque doesn't change, you won't have much speed error to start with.
Second, in some applications, especially those involving web products and tension control, cumulative error is just as important as actual error. For example, even if you are very accurate with actual error, if it is all negative or all positive, eventually you are going to have too much or too little tension. No encoderless system will assure non-cumulative error. For that you need an encoder.
Third, speed reference error is often overlooked. That is error either in the speed signal going into the drive or error in the drive translating the input command into an actual output speed. Usually, the majority of this error is due to the analog input terminal analog-to-digital conversion. A 10 bit resolution A/D input will not be nearly as accurate as a 14 bit resolution input. This is a matter of purchasing a drive with the input resolution adequate for the intended purpose.
Skogsgurra
Electrical
Actually, a system with an encoder will be needed to produce a cumulative error. Any system without an encoder will be so badly controlled that it isn't runnable and you will never experience the cumulative error problem. Talking paper machines and/or sheet rolling mills now.
Actually, again, even the best encoder and speed control system won't work with some materials. A couple of perfect drives with a sheet running between them behave like a perfect integrator - it integrates over time. And if time is long enough, the sheet will either drop or tear (if not resilient).
That's why you find devices like dancing rollers, triangulation sensors or sometimes even ultra-sound position sensors to control hang or force sensors to control force in many sections in a paper machine or rolling mill.
Gunnar Englund
--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...
Actually, again, even the best encoder and speed control system won't work with some materials. A couple of perfect drives with a sheet running between them behave like a perfect integrator - it integrates over time. And if time is long enough, the sheet will either drop or tear (if not resilient).
That's why you find devices like dancing rollers, triangulation sensors or sometimes even ultra-sound position sensors to control hang or force sensors to control force in many sections in a paper machine or rolling mill.
Gunnar Englund
--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...
Motor slip is always negative error (when motoring) and sensorless systems generally are dealing with the minimization of slip error. In my experience, that makes sensorless vector error almost always negative, however small.
I have seen a few systems with persistent positive error when the motor nameplate data was in error, especially full load speed, but that is most uncommon.
Cumulative error definitely happens in sensorless systems and there is no way to really control it or make it cycle around the zero error point.
I have seen a few systems with persistent positive error when the motor nameplate data was in error, especially full load speed, but that is most uncommon.
Cumulative error definitely happens in sensorless systems and there is no way to really control it or make it cycle around the zero error point.
- Status
Similar threads
- Question
- Question
- Locked
- Question